People in Glass Houses Shouldn't File $100M ATMs

Glass House Needs to Get Its House in Order

Glass House Brands Stock: Why $GLAS Fell 70% From Its High

Glass House Brands $GLAS printed a 52-week low of $4.20 today. It traded $4.33 late in the session, down 15.6% on the day on 4.46 million shares. That's more than double its 10-day average. The June 18 high was $13.93. Do the math and you get a 69% drawdown in under four months. The market cap is now $465 million. In June it was north of $1.2 billion.

We've been asked all week what broke. Nothing broke. The stock got priced for a story the numbers never backed up. Then the paper showed up.

Paying at the top and forgetting the valuation is a recipe for disaster

Start with the multiple, because that's where the damage was built in. Glass House uplisted to the NYSE on June 30, three weeks after Trulieve $TRLV became the first plant-touching U.S. operator on a major exchange. The tape loved it. The stock ran from about $10 in late May to almost $14 by June 18.

At the top, Glass House carried a market cap of roughly $1.2 billion on about 89 million shares. Trailing revenue was $177 million, down about 10% year over year. Trailing adjusted EBITDA was negative. Add it up: a $2.3 million loss in Q3 2025, a $3.3 million loss in Q4, a $4.2 million loss in Q1 2026, then a $5.7 million gain in Q2. That nets to about negative $4 million over four quarters. So EV/EBITDA wasn't high. It was undefined. On revenue, the enterprise value sat near 7 times sales.

Now look at the names we call the cream of the crop. Trulieve trades near 6 times trailing EBITDA and about 1.7 times revenue. Green Thumb $GTBIF sits near 6 times EBITDA too, and it bought back 6 million shares for $33 million in Q1 with real cash flow. Curaleaf $CURLF is the mildly more expensive of the big 3, but Curaleaf is also the only one in Europe which is where the next decade of money in cannabis will be made. Glass House was being priced like a software company when It grows weed in Camarillo. We said as much when Trulieve uplisted, and the gap only widened from there.

Even if you gave management full credit for the original 2026 guide, which called for $235 million to $245 million in revenue and adjusted EBITDA in the high $40 million range, the peak price was still 5 times sales and about 26 times forward EBITDA. And that guide didn't survive the year. Q2 came in at $47.0 million in revenue against a $54.4 million consensus. Gross margin was 34%, down from 55% a year earlier. Management shelved the full-year guide, pointing to the retail deconsolidation and open questions on out-of-state sales and hemp. Canaccord trimmed its target to $15 from $16 after the print.

Then the paper showed up, because it always does

Is the company selling stock? Yes, and it isn't hiding it. The ATM program started at $25 million in late 2024. It went to $50 million in May. On July 15, two weeks after the uplist, it went to $100 million, with sales permitted directly on the NYSE. In Q1 alone the company sold 2.09 million shares at an average of $8.98 for $18.7 million gross. Q2 added $4.9 million net, and another $1.6 million went out at $11.29 after the quarter closed. Management calls the ATM a long-term, opportunistic source of capital. Fine. But a $100 million shelf on what is now a $465 million company is a standing overhang. The Q3 print, expected around November 11, will tell us how hard they leaned on it into the decline.

The warrants did more damage than the ATM. In June, with the stock holding above $12 for 10 of 15 sessions, the company accelerated the expiry of its Series B, C and D warrants. Strikes were $5, $5 and $6. Holders had until July 23 and could exercise cashless. All of them did. That put 7.4 million new shares into the float, 7.8 million once you count the legacy warrant cleanup in May. Roughly 9% of the company landed with people sitting on gains of 100% or more, with no lockup, in a stock that traded a few hundred thousand shares a day before the uplist. The share count went from 84.7 million at the end of Q1 to 89.0 million at the end of Q2 to 93.0 million today. That's your July-to-September grind.

And no, converts aren't maturing into this. The only convertible is the Series E preferred at the GH Group subsidiary: $77.5 million face, a 12% coupon, convertible at $9.00 and exchangeable one-for-one into the public shares at any time. Call it 8.6 million potential shares. That paper was in the money from $9 to the high, so it capped every rally above $9. At $4.33 it's underwater and nobody's converting. The Series D preferred, $15 million at 15% stepping to 20% in August 2028, doesn't convert at all. Together the preferreds cost about $2.9 million a quarter in cash dividends. On a company that printed $5.7 million of adjusted EBITDA in its best quarter of the year, that's not nothing.

The house itself looks like a tenement but fixable

This is the part we'd fix first if we sat on this board. On September 29, Glass House announced the planned retirement of CFO Mark Vendetti. He stays through March 31, 2027, and a search is on. Earlier in September, the company's own press releases had already retitled Jon DeCourcey from Vice President of Investor Relations to Vice President of Capital Markets. By the September 29 release, no in-house contact appeared at all. The only name listed was the outside IR firm. From our view who ever is doing outside IR reminds us of the realtor in “The Big Short” that calls this a “gully.” This is a 4 alarm fire from a communication standpoint

We've heard the chatter that DeCourcey was pushed out. We've heard the louder chatter that some holders want Kyle Kazan replaced as CEO. We can't confirm either one, and the company hasn't said a word. But let's be honest about the second one. Nobody's replacing Kazan. The founders hold every one of the company's multiple voting shares, and each one carries 50 votes. Kazan holds about 43% of them and co-founder Graham Farrar about 28%. Between the two of them that's a majority of all the votes in the company, before you count a single equity share they own. A CEO change here runs through a board Kazan chairs and a vote Kazan controls. Short of finding a dead body in his trunk or him breaking the law in some other way Kazan stays.

Two dates matter. The first already happened. At the June 18 annual meeting, the director vote totaled about 16.7 million shares, against 239 million a year earlier. The super-votes don't appear to have been cast in that election, and we don't know why. With the float voting on its own, four directors won with only about 61% support and 39% withheld. That's the float telling you what it thinks of this board. The second date is June 29, 2027. That's when the multiple voting shares sunset and get redeemed for a tenth of a penny each, unless shareholders extend them again. They did that once already, in 2023, with 97% of the equity vote. We doubt it goes that smoothly twice.

There's a third thing every Glass House holder already knows about, so we'll say it plainly. In 2023, Catalyst Cannabis CEO Elliot Lewis accused Glass House, first on Instagram and then in a Los Angeles Superior Court complaint, of moving product into the illicit market through burner distributors. Glass House denied it and countersued for defamation. The cases have bounced around the Los Angeles courts since, and nothing has ever been proven against Glass House. We're not relitigating it here. We raise it because a company that carries that history has to be cleaner than clean on disclosure. Right now it has a departing CFO and no visible IR lead. That's the wrong look.

So the credibility problem is real. It's fixable. But only the people who already run the place can fix it, and the list is short.

Name the CFO successor before the Q3 call, not after. Put a real investor relations function back in place and let it talk. Re-issue 2026 guidance on the Q3 call with numbers that reflect the retail deconsolidation, so the Street can rebuild its models. Hit the sub-$100 per pound production cost management promised for year-end. Either use the ATM in the open or say plainly that it stays in the drawer until the stock recovers. And tell the market now whether management plans to ask for another super-vote extension next June, because that vote is coming either way.

What we're doing with $GLAS is not much of anything

Bearish on the setup, neutral on the business, taking a starter on the risk/reward trade. Glass House still grew a record 246,000 pounds last quarter at $122 a pound. It's the only operator in the country that can talk about a million pounds of annual production with a straight face. That's a real asset. The stock just isn't a momentum vehicle anymore, and it won't be one again until the capital structure stops leaking.

At $4.33 the enterprise value is roughly $540 million once you add back the $92.5 million of preferreds and net $72 million of debt against $90 million of cash. That's about 3 times trailing revenue. Cheaper, not cheap. Trulieve is still at 1.7 times with a real EBITDA line under it. The warrant holders who exercised at $5 and $6 and held are now underwater. The forced selling is probably done. The regret selling has started.

Our standing rule in this sector applies, the same one we laid out in our read on the rescheduling fight: assume the news won't happen, and trade around a core position. We'd rather own the cream of the crop at 6 times EBITDA than catch this knife on a story. If Glass House gets its house in order by the Q3 call, we'll say so, but stop losses are always a good idea in this sector.

Want our cannabis sector work the day it publishes? Subscribe free at StockAlpha.ai.